Polish is quantitatively different on quartzite flakes used on different worked materials [Python analysis]

This upload includes the following files related to the Python analysis:
    1.    Raw data as a XLSX table (processing-quartzite-final-2020-04-29.xlsx) is the output from R Script #1 (see https://doi.org/10.5281/zenodo.3979139), even though the filename is slightly different.

Plus, for each analysis (full and restricted datasets), included in the corresponding ZIP archive:
    2.    Jupyter notebooks of the analysis (Classification_RandSplitFeature_Revision_VXX.ipynb) rendered to HTML file (Classification_RandSplitFeature_Revision_VXX.html)
    3.    Dataframe including the artificially filled datapoints
    4.    Output of the analysis as PDF:
    •    Confusion matrices ("CM")
    •    Decision trees on selected features ("DecisionTreeSel")
    •    Balanced accuracy vs. maximum depth ("depth")
    •    Mutual information ("MI")
    •    Pairplots of selected features ("pairplot")
    •    Performance on training sets ("performance")

Instructions to download all files at once are given here: https://doi.org/10.5281/zenodo.4011952